Discussion Points 
31 
Discussion Points 
DP-1 for Fig. 8 and its data, I. Tolosa: From my point of view, calculating the 
RSD using the raw counts is not correct, because the raw counts will depend 
on the concentration (or dilution) of the sample. I meant that gradual evapora 
tion of the solvent from the vials seating on the GC-auto sampler will cause an 
increase on the raw counts with time (increase in the absolute height value of 
the peak) and therefore normalized values to a fixed compound within the sam 
ple should be taken In account for calculating the RSD. In this sense, I suggest 
to add the same Fig based on normalized counts as it is in new Fig. 8. 
DP-2 for Fig. 8 and its data, I. Tolosa: Drift problems are minimized by normali 
zation to the maximum peak intensity In each individual sample. The higher 
RSD for the high boiling hydrocarbon is mainly due to the mass discrimination 
occurring in the injection port and the use of peak height Instead of peak area. 
DP-3 for Fig. 10 and Fig. 12 and their data, I. Tolosa: Similar comment as 
before; RSD should be based on normalized data and not raw counts. Moreo 
ver, here It Is shown that the peak area provides lower RSD values than the 
peak height, at least for the higher n-alkanes with a high boiling point. 
DP-4 for paragraph 10 In section IRMS data, I. Tolosa: It Is obvious that IRMS 
laboratories participating in the same intercalibration exercise should use the 
same certified standard reference to avoid and minimize this discrepancy. 
Traceability to the same reference Is essential for comparability of the results. 
More details on the calibrated 13C standards used for the measurements Is 
needed to understand the discrepancy of 3 per mil. 
DP-5 for section IRMS data, C. Blaga: Not “very” surprising the discrepancies 
In absolute results as the instruments are not always calibrated In the same 
manner in between labs. Different standards used, columns, lab conditions In 
general. We haven’t been measuring steadily on the GC-IRMS and this gave 
also so stability Issues. The standard mixtures I have mentioned refer to an 
“In-house” made mixture. The alkanes are commercially available In powder 
form and from each a certain amount was taken and dissolved In hexane to 
obtain the mixture used. The powders were previously measured with EA-IRMS 
to have an initial value for the Isotopes and further the mix analyzed on the 
GC-IRMS (with the expected off-set, and less expected variability In time due 
to the low frequency In use of the instrument). 
DP-6 for section IRMS data, C. Blaga: As a reference PDB is exhausted hence 
only VPDB is available. During each analysis, at least two pulses of reference 
C0 2 gas from the cylinder are introduced. This C0 2 Is previously calibrated 
against other VPDB calibrated international standards, themselves calibrated 
against international IAEA standards. The reference C0 2 gas can be addition 
ally calibrated against in-house standards but as such standards are not exist 
ent and frequently tested among labs this has not been done In our case.